JP7389278B2 - Battery diagnostic device, battery pack, battery system and battery diagnostic method - Google Patents

Description

The present invention relates to a technique for detecting an abnormality in a battery cell.

This application claims priority based on Korean Patent Application No. 10-2020-0096786 filed on August 3, 2020, and all contents disclosed in the specification and drawings of the corresponding application are incorporated into this application. .

Recently, the demand for portable electronic products such as notebook PCs, video cameras, mobile phones, etc. has rapidly increased, and as the development of electric vehicles, energy storage batteries, robots, satellites, etc. Research into high-performance batteries that are capable of

Currently, commercially available batteries include nickel-cadmium batteries, nickel-metal hydride batteries, nickel-zinc batteries, and lithium batteries. Among these, lithium batteries have almost no memory effect compared to nickel-based batteries, and they charge and discharge easily. It is attracting attention because of its advantages of free self-discharge, extremely low self-discharge rate, and high energy density.

Recently, as applications requiring high voltage have become widespread, a structure in which a plurality of battery cells are connected in series within a battery pack has become widely used. The greater the number of battery cells included in a battery pack, the greater the possibility that battery cell abnormalities will occur. Therefore, the need for diagnostic technology that can accurately detect abnormalities in battery cells is increasing.

Conventionally, cell information including multiple parameters related to the state of battery cells (e.g., voltage, current, temperature) is monitored, and the usage state of the battery cell (e.g., charging, discharging, resting) and the monitored cell information are monitored. A method is used to detect whether there is an abnormality in the battery cell based on the following.

However, the above abnormality detection method requires a process in which the battery management system monitors the cell information of the battery cells using various sensors, so the amount of calculation required for abnormality detection becomes excessive and takes a long time. It takes. In particular, in a structure in which the battery management system is supplied with power from a battery cell, the electrical energy of the battery cell may be continuously consumed during the operation of the battery management system for abnormality detection.

Furthermore, conventionally, abnormality in a battery cell has been detected based on a sudden change in cell information of the battery cell over a short period of time. However, the cell information of an abnormal battery cell does not necessarily change rapidly in a short period of time, but may change gradually over a long period of time, so there is a possibility that an abnormality in a battery cell cannot be detected at an appropriate time.

The present invention has been made in view of the above problems, and provides a battery diagnostic device, a battery pack, and a battery that use the cell voltage of each of a plurality of battery cells connected in series as a single parameter for abnormality detection. The purpose is to provide a system and battery diagnosis method.

Further, the present invention generates an observation matrix, which is a data set including a plurality of observation voltage vectors indicating the voltage history (time series) of cell voltages of each of a plurality of battery cells observed in the same period, and then generates an observation matrix. Another object of the present invention is to provide a battery diagnostic device, a battery pack, a battery system, and a battery diagnostic method for detecting battery cell abnormalities by identifying abnormal behavior of the cell voltage of each battery cell based on analysis results. do.

Other objects and advantages of the present invention can be understood from the following description and will be more clearly understood from the embodiments of the invention. Further, the objects and advantages of the present invention can be realized by means and combinations thereof shown in the claims.

A battery diagnostic device according to one aspect of the present invention includes a memory configured to store an observation matrix including a plurality of observed voltage vectors indicating a time series of cell voltages of each of a plurality of battery cells; a controller configured to determine a principal component vector, a plurality of singular values, and a plurality of coefficient vectors. Each coefficient vector includes a plurality of coefficients that correspond one-to-one to the plurality of observed voltage vectors. For each coefficient vector, the control unit compares a plurality of coefficients included in the coefficient vector, determines an ineffective coefficient among the plurality of coefficients, and determines an ineffective coefficient among the plurality of principal component vectors. Based on the principal component vector corresponding to the vector, the singular value corresponding to the coefficient vector among the plurality of singular values, and the ineffective coefficient, detect an abnormality in the battery cell corresponding to the ineffective coefficient among the plurality of battery cells. configured to detect.

The control unit may be configured to apply a matrix decomposition algorithm to the observation matrix and determine a first sub-matrix, a second sub-matrix, and a third sub-matrix. The first sub-matrix includes the plurality of principal component vectors as column vectors. The second sub-matrix includes the plurality of singular values as main diagonal components. The third sub-matrix includes the plurality of coefficient vectors as row vectors.

The control unit may be configured to determine, among the plurality of coefficients, a coefficient whose absolute value of a difference from an average of the plurality of coefficients is larger than a first reference value as the ineffective coefficient.

The control unit may be configured to determine the first reference value to be equal to a value obtained by multiplying a standard deviation of the plurality of coefficients by a first scaling factor.

The control unit controls, for each coefficient vector, a principal component vector corresponding to the coefficient vector among the plurality of principal component vectors, a singular value corresponding to the coefficient vector among the plurality of singular values, and the ineffective coefficient. and extracts a partial voltage vector of the observed voltage vector corresponding to the ineffective coefficient among the plurality of observed voltage vectors. The battery cell may be configured to detect a battery cell corresponding to the ineffective coefficient among the battery cells as abnormal.

The control unit may be configured to determine the voltage characteristic value equal to a difference between a maximum partial voltage and a minimum partial voltage among a plurality of partial voltages included in the partial voltage vector.

The control unit may be configured to determine the second reference value to be equal to a value obtained by multiplying the voltage resolution of the voltage measurement circuit by a second scaling factor.

The control unit may be configured to output a fault message when a ratio of a maximum singular value to a minimum singular value among the plurality of singular values is less than a set value.

A battery pack according to another aspect of the present invention includes the battery diagnostic device.

A battery system according to yet another aspect of the present invention includes the battery pack.

A battery diagnosis method according to still another aspect of the present invention provides a battery diagnosis method in which a plurality of principal component vectors, a plurality of singular values, and a plurality of including determining a coefficient vector. Each coefficient vector includes a plurality of coefficients that correspond one-to-one to the plurality of observed voltage vectors. The battery diagnosis method includes, for each coefficient vector, comparing a plurality of coefficients included in the coefficient vector to determine an ineffective coefficient among the plurality of coefficients; Based on the principal component vector corresponding to the coefficient vector, the singular value corresponding to the coefficient vector among the plurality of singular values, and the ineffective coefficient, calculate the battery cell corresponding to the ineffective coefficient among the plurality of battery cells. The method further includes detecting an anomaly.

The step of determining an ineffective coefficient among the plurality of coefficients includes determining, among the plurality of coefficients, a coefficient whose absolute value of the difference from the average of the plurality of coefficients is larger than a first reference value as the ineffective coefficient. It may be at the stage of

The step of detecting an abnormality in a battery cell corresponding to the ineffective coefficient among the plurality of battery cells includes detecting a principal component vector corresponding to the coefficient vector among the plurality of principal component vectors and a corresponding one among the plurality of singular values. multiplying a coefficient vector by a corresponding singular value and the non-effective coefficient to extract a partial voltage vector of the observed voltage vector corresponding to the non-effective coefficient among the plurality of observed voltage vectors; and voltage characteristics of the partial voltage vector. The method may include detecting a battery cell corresponding to the ineffectiveness coefficient among the plurality of battery cells as abnormal if the value is larger than a second reference value.

According to at least one embodiment of the present invention, in detecting an abnormality in each of a plurality of battery cells connected in series, only the cell voltage excluding current and temperature is used, which is necessary for abnormality detection. The amount of calculation, time and power can be reduced.

Further, according to at least one embodiment of the present invention, the observation is a data set including a plurality of observed voltage vectors indicating the voltage history (time series) of each cell voltage of a plurality of battery cells observed in the same period. After the matrix is generated, abnormal behavior of the cell voltage of each battery cell is identified based on the analysis result of the observed matrix, thereby improving the accuracy of abnormality detection for the battery cells.

The effects of the present invention are not limited to the above-mentioned effects, and other effects of the present invention not mentioned will be clearly understood by those skilled in the art from the description of the claims.

The following drawings attached to this specification illustrate preferred embodiments of the present invention, and together with the detailed description of the invention serve to further understand the technical idea of the invention, The interpretation shall not be limited to only the matters shown in the drawings.

1 is a diagram showing the configuration of a battery system according to the present invention. It is a graph showing a change in cell voltage of a battery cell over time. 3 is a diagram for explaining an observation matrix as a data set showing the voltage history of the battery cell shown in FIG. 2. FIG. It is a graph showing a coefficient vector. FIG. 3 is a diagram for explaining the relationship between an ineffective coefficient and a partial voltage vector. 1 is a flowchart illustrating a battery diagnosis method according to a first embodiment of the present invention. 7 is a flowchart illustrating a battery diagnosis method according to a second embodiment of the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. Prior to this, the terms and words used in this specification and claims should not be interpreted to be limited to their ordinary or dictionary meanings, and the inventors themselves have expressed their intention to explain the invention in the best way possible. Therefore, the meaning and concept of the term should be interpreted in accordance with the technical idea of the present invention, based on the principle that the concept of the term can be appropriately defined.

Therefore, the embodiments described in this specification and the configurations shown in the drawings are only one of the most desirable embodiments of the present invention, and do not represent the entire technical idea of the present invention. It is to be understood that there may be various equivalents and modifications that may be substituted for these at the time of filing.

Terms including ordinal numbers, such as first, second, etc., are used to distinguish one of the various components from the rest, and the components are not limited by these terms.

Furthermore, throughout the specification, when a certain part is said to "include" a certain component, unless there is a statement to the contrary, this does not mean that other components are excluded, but it may further include other components. It means that. Furthermore, terms such as "control unit" described in the specification indicate a unit that processes at least one function or operation, and this may be realized by hardware, software, or a combination of hardware and software. .

Furthermore, throughout the specification, when a certain part is "connected (connected)" to another part, this does not only mean that it is "directly connected (connected)"; This also includes cases in which they are "indirectly connected (connected)" through other elements in between.

FIG. 1 is a diagram showing the configuration of a battery system according to the present invention.

An energy storage system is shown in FIG. 1 as an example of a battery system 1 . Referring to FIG. 1, a battery system 1 includes a battery pack 10 and a switch 20. The battery system 1 may further include at least one of a remote controller 240 and a power conversion system 30. The battery system 1 is not limited to an energy storage system, and may take any form as long as it has a charging function and/or a discharging function for the battery pack 10 installed therein, such as an electric vehicle or battery inspection equipment. .

The battery pack 10 includes a positive terminal P+, a negative terminal P-, a cell group 11, and a battery management system 100. Cell group 11 includes a plurality of battery cells BC ₁ to BC _n (n is a natural number of 2 or more) electrically connected between positive terminal P+ and negative terminal P-. In FIG. 1, a plurality of battery cells BC ₁ to BC _n are shown connected in series within a cell group 11. As shown in FIG. Hereinafter, when describing common contents for a plurality of battery cells BC ₁ to BC _n , reference numeral "BC" will be used as a battery cell.

A positive terminal and a negative terminal of a battery cell BC are electrically connected to other battery cells BC via a conductor such as a bus bar. Battery cell BC may be a lithium ion battery cell. Of course, the type of battery cell BC is not particularly limited as long as it can be repeatedly charged and discharged.

A switch 20 is provided on the power line PL for the battery pack 10. While the switch 20 is turned on, power can be transferred from one of the battery pack 10 and the power conversion system 30 to the other. The switch 20 may be implemented using any one or a combination of two or more known switching devices such as a relay, a field effect transistor (FET), and the like.

A power conversion system 30 is operably coupled to one of the battery management system 100 and the remote controller 240. When two structures are operably coupled, it means that the two structures are directly or indirectly connected so that they can transmit and receive signals in either one or both directions. Power conversion system 30 may generate DC power for charging cell group 11 from AC power supplied by electrical system 40 . Power conversion system 30 may generate alternating current power from direct current power from battery pack 10.

Battery management system 100 may include voltage measurement circuit 110 and battery controller 140. The battery management system 100 may further include at least one of a current sensor 120, a temperature sensor 130, and an interface unit 150. The interface unit 150 may be included in the battery controller 140.

The voltage measurement circuit 110 is provided so as to be electrically connectable to a positive terminal and a negative terminal of the battery cell BC. Voltage measurement circuit 110 may measure a cell voltage, which is the voltage applied across battery cell BC, and output a signal indicating the measured cell voltage to battery controller 140.

Current sensor 120 is electrically connected in series to cell group 11 by power line PL. For example, a shunt resistor, a Hall element, or the like can be used as the current sensor 120. Current sensor 120 may measure the current flowing through cell group 11 and output a signal to battery controller 140 indicative of the measured current.

Temperature sensor 130 is placed in an area within a predetermined distance from cell group 11 . For example, a thermocouple or the like may be used as the temperature sensor 130. Temperature sensor 130 may measure the temperature of cell group 11 and output a signal indicative of the measured temperature to battery controller 140.

Battery controller 140 is operably coupled to voltage measurement circuit 110, current sensor 120, temperature sensor 130, and/or interface unit 150. At least one of the battery controller 140 and the remote controller 240 may turn on/off the switch 20 according to the diagnosis result for the cell group 11 .

The interface unit 150 may be communicatively coupled to the remote controller 240 of the battery system 1 . The interface unit 150 may transmit signals from the remote controller 240 to the battery controller 140 and transmit signals from the battery controller 140 to the remote controller 240. The signal from battery controller 140 may include information for notifying an abnormality in battery cell BC. Communication between the interface unit 150 and the remote controller 240 may be performed using, for example, a LAN (local area network), a CAN (controller area network), a wired network such as a daisy chain, and/or Bluetooth (registered trademark), ZigBee, or Wi-Fi. - Short-range wireless networks such as Fi® may be utilized. The interface unit 150 may include an output device (eg, a display, a speaker) that provides information received from the battery controller 140 and/or the remote controller 240 in a user-perceivable form. The remote controller 240 controls the battery pack 10, the switch 20, and the power conversion system based on cell information (for example, cell voltage, current, temperature, SOC, abnormality of the battery cell BC) collected through communication with the battery management system 100. At least one of 30 can be controlled.

Battery controller 140 includes memory 141 and control section 142. Remote controller 240 may include memory 241 and control unit 242 . Remote controller 240 may further include communication circuitry 243. Remote controller 240 may be implemented in the form of a cloud server or mobile diagnostic equipment. The communication circuit 243 is for wired/wireless communication with the battery management system 100.

The control unit 142 and the control unit 242 each include an ASIC, a DSP (digital signal processor), a DSPD (digital signal processing device), and a PLD (programmable locomotive device) in terms of hardware. gic device, programmable logic device ), a field programmable gate array (FPGA), a microprocessor, and other electrical units for performing functions.

At least one of the memory 141 and the memory 241 may store programs and various data necessary for executing a battery diagnosis method (diagnosis stage) according to an embodiment described later. The memory 141 and the memory 241 are each of, for example, a flash memory (registered trademark) type, a hard disk type, an SSD type (Solid State Disk type), or an SDD type (Silicon Disk type). Drive type, silicon disk drive type), multimedia card micro type, RAM (random access memory), SRAM (static random access memory), ROM (read-only) memory, read-only memory), EEPROM (electrically erasable programmable read-only memory), PROM (programmable read-only memory) y, programmable read-only memory). obtain. At least one of the memory 141 and the memory 241 may store data and algorithms necessary for detecting an abnormality in the battery BC by performing a diagnostic step (FIGS. 2 to 6), which will be described later. The memory 141 and the control unit 142 may be integrated in the form of a single chip. The memory 241 and the control unit 242 may be integrated in the form of a single chip.

Battery controller 140 is one example of a battery diagnostic device according to the present invention, and remote controller 240 is another example of a battery diagnostic device according to the present invention. That is, the diagnostic steps described below with reference to FIGS. 2 to 7 are performed by at least one of the battery controller 140 and the remote controller 240, which are provided as battery diagnostic devices.

The battery diagnostic device according to the present invention may perform a diagnostic step (see FIGS. 6 and 7) to detect abnormalities in a plurality of battery cells BC ₁ to BC _n . The diagnosis stage includes voltage data acquired during a specific period (for example, a certain period of time in the past) during which the cell group 11 is maintained in a predetermined diagnosable state (for example, rest, constant current charging, constant voltage charging, constant current discharging). (See X ₁ to X _n in FIG. 2). The following description will be made assuming that the battery controller 140 is provided as a battery diagnostic device.

FIG. 2 is a graph showing changes over time in the cell voltage of a battery cell, and FIG. 3 is a diagram for explaining an exemplary observation matrix as a data set showing the voltage history of the battery cell shown in FIG. It is.

The control unit 142 determines the voltage value of each cell voltage of the plurality of battery cells BC ₁ to BC _n at each set time based on the voltage signal from the voltage measurement circuit 110, and stores the determined voltage value in the memory 141. can be recorded. The set time may be the same as the length of a cycle for abnormality detection, which will be described later.

The control unit 142 determines an observation matrix X that includes a plurality of observed voltage vectors X ₁ to X _n for a specific period Δt during a certain period of time in the past. A moving window 200 may be used to determine the observation matrix X. In one example, the plurality of observed voltage vectors X ₁ to X _n indicate changes over time in the cell voltages of each of the plurality of battery cells BC ₁ to BC _n measured at each set time within the moving window 200. The moving window 200 is used to determine a period Δt during which a plurality of observed voltage vectors X ₁ to X _n are acquired while moving by a set time every set time. The size Δt of the moving window 200 may be predetermined or may be adjusted by the controller 142.

The cell voltage of the battery cell BC is measured multiple times in time series by the voltage measurement circuit 110 (for example, m times in total, where m is a natural number of 2 or more), and the measured value of the cell voltage is measured by the control unit 142. It can be recorded in memory 141. For example, when the size of the moving window 200 = 200 seconds and the set time = 1 second, since m = 200, the cell voltage of the battery cell BC is measured 200 times within the moving window 200.

Referring to FIG. 2, a curve 210 shows a change over time in the cell voltage of the k-th battery cell BC _k among the plurality of battery cells BC ₁ to BC _n . The abnormal condition may be a condition that induces abnormal behavior of the cell voltage, such as an internal short circuit. t ₁ and t _m in FIG. 2 are the start and end times of the specific period Δt, respectively, and t _i is the time corresponding to the time index i within the specific period Δt. In the k-th battery cell BC _k , k is a natural number less than or equal to n, and may represent a cell index used to distinguish among the plurality of battery cells BC ₁ to BC _n .

Hereinafter, the abnormality detection operation according to the present invention will be described with reference to the k-th battery cell BC _k . The explanation regarding the k-th battery cell BC _k is also commonly applicable to the remaining battery cells BC of the plurality of battery cells BC ₁ to BC _n .

Referring to FIG. 3, the observation matrix X is an m×n matrix including m rows and n columns. In the following, for convenience of explanation, it is assumed that m is a natural number greater than n, i is a natural number from 1 to m, j is a natural number from 1 to n, and k is a natural number less than n.

The n column vectors of the observation matrix X can correspond one-to-one to the plurality of observation voltage vectors X ₁ to X _n . That is, each of the plurality of observation voltage vectors X ₁ to X _n is a column vector of the observation matrix X, and includes m elements (measured values of cell voltage). The k-th observed voltage vector X _k is the cell voltage of the k-th battery cell BC _k measured m times arranged in time order, that is, the measured cell voltage of the k-th battery cell BC _k x _1k to x _mk is a time series (set) of The k-th observed voltage vector X _k may be the k-th column vector of the observation matrix X. Referring to _{FIG .} 2, in the observation matrix elements (which may be referred to as "data" or "components").

The control unit 142 may apply matrix decomposition to the observation matrix X to extract the first sub-matrix A, the second sub-matrix B, and the third sub-matrix C ^T from the observation matrix X. That is, the observation matrix X can be decomposed into a first sub-matrix A, a second sub-matrix B, and a third sub-matrix ^CT . Examples of algorithms used for matrix decomposition include singular value decomposition (SVD) and principal component analysis (PCA). In this specification, the symbol "T" written in a superscript on the right side of a matrix means a transposed matrix. As shown, the product of the first sub-matrix A, the second sub-matrix B, and the third sub-matrix ^CT is the same as the observation matrix X.

The first sub-matrix A is an m×m matrix. The second sub-matrix B is an m×n matrix. The third sub-matrix ^CT is an n×n matrix.

The first sub-matrix A is an orthogonal matrix and includes a plurality of principal component vectors A ₁ to A _m . Each principal component vector of the plurality of principal component vectors A ₁ -A _m may be referred to as a "left singular vector." Each principal component vector includes m elements and may be a column vector of the first sub-matrix A. That is, the first sub-matrix A is expressed as follows.
A=[A ₁ A ₂ ...A _m ]
Ai=[a _1i a _2i ... _ami ] ^T

Among the plurality of principal component vectors A ₁ to A _m , principal component vectors A ₁ to A _n indicate dispersion information of the observation matrix X. The j-th principal component vector A _j corresponds to the direction of the axis where the variance of the elements of the observation matrix X is the j-th largest. That is, when the elements of the observation matrix X are mapped once on each axis of the plurality of principal component vectors A _{1 to A} big.

The larger the magnitude of the variance of the j-th principal component vector A _j , the greater the explanatory power of the j-th principal component vector A _j for the distribution state of the elements of the observation matrix X. The greater the explanatory power of the j-th principal component vector A _j is, the more the j-th principal component vector A _j has common voltage behavior characteristics (for example, normal voltage behavior) of a plurality of battery cells BC ₁ to BC _n within the moving window 200. (trends). Conversely, the smaller the magnitude of the variance of the j-th principal component vector A _j , the weaker the explanatory power of the j-th principal component vector _{A j becomes} . ).

The second sub-matrix B is a diagonal matrix and includes a plurality of singular values b ₁₁ to b _nn as elements on the main diagonal. That is, the second sub-matrix B is expressed as follows.
B=[B ₁ B ₂ ...B _n ]
B _j = [b _1j b _2j ... b _mj ] ^T
If i≠j, b _ij is 0. b _jj is the j-th singular value.

That is, among the total m×n elements of the second sub-matrix B, all the remaining elements except for n elements b ₁₁ to b _nn on the main diagonal are 0. Accordingly, principal component vectors A _n+1 to _m among the plurality of principal component vectors A ₁ to A _m are redundant in explaining the dispersion information of the observation vector X.

The plurality of singular values b ₁₁ to b _nn can satisfy the following relationship.
b ₁₁ ≧b ₂₂ ≧…≧b _nn ≧0

The plurality of singular values b ₁₁ to b _nn are referred to as the first to nth singular values in descending order of magnitude, and b _jj may be the jth largest singular value among the plurality of singular values b ₁₁ to b _nn . .

The plurality of singular values b ₁₁ to b _nn indicate explanatory power information of the plurality of principal component vectors A ₁ to A _n . The singular value b _jj of the second sub-matrix B indicates the explanatory power of the j-th principal component vector A _j .

The third sub-matrix C ^T is an orthogonal matrix and includes a plurality of coefficient vectors C ₁ ^T to C _n ^T. Each coefficient vector of the plurality of coefficient vectors C ₁ ^T -C _n ^T may be referred to as a "right singular vector." Each coefficient vector includes n components and may be a row vector of the third sub-matrix ^CT . The third sub-matrix C ^T is expressed as follows.
C ^T = [C ₁ C ₂ ... C _n ] ^T = [C ₁ ^T ; C ₂ ^T ;...; C _n ^T ]
C _j ^T = [c _j1 c _j2 ... c _jn ]

The plurality of coefficient vectors C ₁ ^T to C _n ^T indicate dependence information of the plurality of observed voltage vectors X ₁ to X _n on the plurality of principal component vectors A ₁ to A _n . Specifically, how significantly each of the plurality of observed voltage vectors X ₁ to X _n is influenced by the j-th principal component vector A _j is determined by the j-th coefficient vector C _j ^T. The j-th coefficient vector C _j ^T includes a plurality of coefficients c _j1 to c _jn that correspond one-to-one to the first to n-th observed voltage vectors X ₁ to X _n . For example, c _jk of the j-th coefficient vector C _j ^T indicates the influence of the j-th principal component vector A _j on the k-th observed voltage vector X _k .

The first to n-th principal component vectors A ₁ to A _n , the first to n-th singular values b ₁₁ to b _nn , and the first to n-th coefficient vectors C ₁ ^T to C _n ^T may have a one-to-one correspondence with each other. .

The observation matrix X is the same as the product of the first sub-matrix A, the second sub-matrix B, and the third sub-matrix ^CT , and can satisfy the relationship according to Equation 1 below.

In Equation 1, A _j is treated as a (m×1) matrix, and C _j ^T is treated as a (1×n) matrix.

Referring to _Equation 1, the _k - _th observed voltage vector The relationship according to Formula 2 is satisfied.

In Equation 2, Y _kj =(b _jj ×A _j ×c _jk ) is the j-th partial voltage vector of the k-th observed voltage vector X _k . The j-th partial voltage vector Y _{kj of the k-th observed voltage vector X k is a voltage component of the k-th observed voltage vector X k that depends} on the j-th principal component vector A _j , and It can be the same as the product of the singular value b _jj and the coefficient c _jk . That is, the j-th partial voltage vector Y _kj restores (approximates) the k-th observed voltage vector X _k using only the j-th principal component vector A _j of the first to n-th principal component vectors A ₁ to A _n. ) may be the result. As a result, the j-th partial voltage vector Y _kj has m elements that correspond one-to-one to the m elements of the k-th observed voltage vector X _k . The elements of each partial voltage vector are sometimes referred to as "partial voltages (or approximate voltages)," and the partial voltage vectors are sometimes referred to as "restored voltage vectors."

The control unit 142 calculates the first to nth principal component vectors A ₁ to A _n , the first to nth singular values b ₁₁ to b _nn , and the first to nth coefficient vectors C ₁ ^T to C _n ^T. Before detecting abnormalities in the first to n-th battery cells BC ₁ to BC _n , the ratio of the maximum singular value b ₁₁ to the minimum singular value _{b nn among} the first to n-th singular values b ₁₁ to b nn can be calculated. When the ratio of the maximum singular value b ₁₁ to the minimum singular value b _nn is less than a predetermined setting ratio (for example, 200%), the control unit 142 sends a fault message indicating that abnormality detection for the battery cell BC is impossible. can be output. Anomaly detection is impossible in a situation where the difference in explanatory power between the plurality of principal component vectors A ₁ to A _n is not clear. That is, in a situation where abnormality detection is impossible, none of the plurality of principal component vectors A ₁ to A _n includes sufficient information about the common voltage behavior characteristics of the plurality of battery cells BC ₁ to BC _n . The reason why abnormality detection is not possible is, for example, a malfunction of the voltage measurement circuit 110 or the number of battery cells BC exceeding a certain percentage (for example, 50%) of the first to nth battery cells BC ₁ to BC _n . This could be due to an abnormality occurring.

If the ratio of the maximum value b ₁₁ to the minimum value b _nn is less than a predetermined setting ratio, the control unit 142 may increase the size of the moving window 200 by a predetermined time in the next cycle. The reason for increasing the size of the moving window 200 is to sufficiently reflect the common voltage behavior characteristics of the plurality of battery cells BC ₁ to BC _n in the observation vector X.

FIG. 4 is a graph showing coefficient vectors. The horizontal axis in FIG. 4 indicates the cell index corresponding to each coefficient of the j-th coefficient vector C _j ^T , and the vertical axis indicates the size of each coefficient. For example, cell index=1 corresponds to the first battery cell _BC1 .

The control unit 142 compares the first to n-th coefficients c _j1 to c _jn included in the j-th coefficient vector C _j ^T and determines whether an ineffective coefficient exists among the first to n-th coefficients c _j1 to c _jn. Decide whether or not to do so. The ineffective coefficient of the j-th coefficient vector C _j ^T indicates the degree of abnormal voltage behavior due to the j-th principal component vector A _j , which is reflected in the voltage history of a specific battery cell corresponding to the ineffective coefficient. . The remaining coefficients excluding the ineffective coefficients can be said to be effective coefficients.

Referring to FIG. 4, the control unit 142 may determine the average c _{j_av} and standard deviation of the first to n-th coefficients c _j1 to c _jn included in the j-th coefficient vector C _j ^T. The control unit 142 may determine the first reference value R _j1 based on the standard deviation of the first to nth coefficients c _j1 to c _jn . For example, the controller 142 may determine the first reference value R _j1 to be equal to the product of the standard deviation and the first scaling factor. The first scaling factor may be recorded in memory 141 in advance.

The control unit 142 selects each coefficient among the first to n-th coefficients c _j1 to c _jn for which the absolute value of the difference from the average c _{j_av} is larger than the first reference value R _j1 as an ineffective member of the j-th coefficient vector C _j ^T. The coefficient can be determined. In FIG. 4, the coefficient c _jk is shown to be smaller than the average c _{j_av} and further smaller than the first reference value R _j1 . Therefore, the control unit 142 may determine the coefficient c _jk to be an ineffective coefficient of the j-th coefficient vector C _j ^T.

FIG. 5 is a diagram for explaining the relationship between the ineffective coefficient and the partial voltage vector. FIG. 5 is a graph showing the j-th partial voltage vector Y _kj of the k-th observed voltage vector X _k corresponding to the ineffective coefficient c _jk of FIG. The j-th partial voltage vector Y _kj is an (m×1) matrix similarly to the k-th observed voltage vector X _k . The horizontal axis in FIG. 5 indicates the time index within the moving window 200, and the vertical axis indicates the partial voltage.

Referring to FIG. 5, the control unit 142 may determine the voltage characteristic value of the partial voltage vector Y _kj based on m partial voltages included in the partial voltage vector Y _kj . The voltage characteristic value may be a parameter indicating the influence (occupancy degree) of the partial voltage vector Y _kj on the k-th observed voltage vector X _k . The fact that the width and/or slope of the voltage change indicated by the partial voltage vector _Ykj within the moving window 200 is large means that the abnormal voltage behavior related to the ineffective coefficient _cjk is large within the k-th observed voltage vector _Xk . It can mean that it is reflected. For example, the control unit 142 may determine the voltage characteristic value to be the same as the difference Δy _kj between the maximum partial voltage y _{kj_max} and the minimum partial voltage y _{kj_min} among m partial voltages of the partial voltage vector Y _kj . Alternatively, the control unit 142 may determine the voltage characteristic value to be the same as the slope of the maximum partial voltage y _{kj_max} and the minimum partial voltage y _{kj_min} .

If the voltage characteristic value of the partial voltage vector Y _kj is greater than the second reference value, the control unit 142 may detect the k-th battery cell BC _k corresponding to the ineffective coefficient c _jk as abnormal. The control unit 142 may determine the second reference value based on the voltage resolution of the voltage measurement circuit 110. In one example, the controller 142 may determine the second reference value based on the product of the voltage resolution and the second scaling factor. The second scaling factor may be pre-recorded in memory. Alternatively, the second reference value may be predetermined to 10.0 mV, etc., taking voltage resolution into consideration. The second reference value is for preventing a normal battery cell from being detected as abnormal due to a measurement error in the cell voltage measured by the voltage measurement circuit 110. If the voltage characteristic value Δy _kj is larger than the second reference value, the k-th battery cell BC _k may be determined to be abnormal.

FIG. 6 is a flowchart illustrating a battery diagnosis method according to the first embodiment of the present invention. The method of FIG. 6 may be repeated at set times.

Referring to FIGS _. 1 to 6, _in step S610, the control unit ₁₄₂ determines an _observation matrix do. The plurality of observed voltage vectors X ₁ to X _n indicate the time series of cell voltages of each of the plurality of battery cells BC ₁ to BC _n measured multiple times (m) in time series within the moving window 200.

In step S620, the control unit 142 determines a plurality of principal component vectors A ₁ to A _n , a plurality of singular values b ₁₁ to b _nn , and a plurality of coefficient vectors C ₁ ^T to C _n ^T from the observation matrix X ( (see formula 1).

Steps S630 to S670 may be performed once for at least one coefficient vector among the plurality of coefficient vectors C ₁ ^T to C _n ^T. For example, steps S630 to S670 may be performed on a predetermined number of coefficient vectors that respectively correspond to a predetermined number of singular values in descending order of magnitude among the plurality of singular values b ₁₁ to b _nn . In another example, steps S630 to S670 may be performed on coefficient vectors corresponding to each singular value whose ratio of the plurality of singular values b ₁₁ to b _nn to the total is less than or equal to a predetermined value.

In step S630, the control unit 142 determines a first reference value R _j ¹ by comparing the plurality of coefficients c _j1 to c _jn with respect to the coefficient vector C _j ^T. Alternatively, the first reference value R _j1 may be a predetermined constant, in which case step S630 may be omitted.

In step S640, the control unit 142 determines whether at least one of the plurality of coefficients c _j1 to c _jn is greater than the first reference value R _j1 . If the value of step S640 is "no", the method ends. If the value of step S640 is "yes", the method proceeds to step S650.

In step S650, the control unit 142 determines a coefficient c _jk larger than the first reference value R _j1 among the plurality of coefficients c _j1 to c _jn as an ineffective coefficient of the coefficient vector C _j ^T.

In step S660, the control unit 142 extracts a partial voltage vector _Ykj of the observed voltage vector _Xk corresponding to the ineffective coefficient _cjk based on the principal component vector _Aj , the singular value _bjj , and the ineffective coefficient _cjk . (See Equation 2).

In step S670, the controller 142 determines a voltage characteristic value Δy _kj of the partial voltage vector Y _kj .

In step S680, the control unit 142 determines whether the voltage characteristic value Δy _kj is greater than a second reference value. If the value of step S680 is "no", the method ends. If the value of step S680 is "yes", it indicates that the battery cell BC _k corresponding to the ineffective coefficient c _jk is detected as abnormal. If the value of step S680 is "yes", the method proceeds to step S690.

In step S690, the controller 142 activates a predetermined protection operation. In one example, the controller 142 turns off the switch 20. In another example, the control unit 142 outputs a diagnostic message indicating information (eg, cell index) of the battery cell BC _k detected as abnormal. Diagnostic messages may be transmitted and received between the battery controller 140 and the remote controller 240 through the interface unit 150. The interface unit 150 outputs visual and/or auditory information corresponding to the diagnostic message.

FIG. 7 is a flowchart illustrating a battery diagnosis method according to a second embodiment of the present invention. The method of FIG. 7 may be repeated at set times.

In the method of FIG. 7, steps S710 to S790 are the same as steps S610 to S690 of FIG. 6, so repeated explanations will be omitted.

The method of FIG. 7 differs from the method of FIG. 6 in that it further includes steps S722 and S724.

In step S722, the control unit 142 determines whether the maximum ratio of the plurality of singular values b ₁₁ to b _nn is greater than or equal to a set ratio. The maximum ratio is the ratio of the maximum value b 11 to the minimum value b _nn among the plurality _of singular values b ₁₁ to b _nn . A value of "No" in step S722 means that there is no principal component vector among the plurality of principal component vectors A ₁ to A _n that has sufficiently greater explanatory power than the remaining principal component vectors. If the value of step S722 is "no", the method proceeds to step S724. If the value of step S722 is "yes", the method proceeds to step S730.

In step S724, the control unit 142 outputs a fault message. A fault message indicates that abnormality detection is not possible. Fault messages may be transmitted and received between the battery controller 140 and the remote controller 240 through the interface unit 150. The interface unit 150 may output visual and/or audio information corresponding to the fault message.

Although the above description with reference to FIGS. 2 to 7 is based on the assumption that the battery controller 140 is provided as a battery diagnostic device, the remote controller 240 instead of the battery controller 140 performs the functions of the battery diagnostic device. I can be in charge. That is, the explanations for each of the control section 142 and the memory 141 are common to the control section 242 and the memory 241. If the remote controller 240 is provided as a battery diagnostic device, the battery management system 100 may transmit data indicating the plurality of observed voltage vectors X ₁ to X _n to the communication circuit 243 of the remote controller 240 through the interface unit 150.

The embodiments of the present invention described above are not necessarily implemented through apparatuses and methods, but may be implemented through programs that implement functions corresponding to the configurations of the embodiments of the present invention or recording media on which the programs are recorded. In addition, such an embodiment should be easily realized by a person who is an expert in the technical field to which the present invention pertains from the description of the above-mentioned embodiments.

Although the present invention has been described above with reference to limited embodiments and drawings, the present invention is not limited thereto. It goes without saying that various modifications and variations are possible within the equivalent range.

Further, the present invention described above is capable of various substitutions, modifications, and changes by persons having ordinary knowledge in the technical field to which the present invention pertains without departing from the technical idea of the present invention. The embodiments are not limited by the accompanying drawings, and all or part of each embodiment can be selectively combined to make various modifications.

1 Battery system 10 Battery pack 11 Cell group 20 Switch 30 Power conversion system
40 Electrical system 100 Battery management system 110 Voltage measurement circuit 120 Current sensor 130 Temperature sensor 140 Battery controller 141 Memory 142 Control section 150 Interface section 200 Moving window 210 Curve 240 Remote controller 241 Memory 242 Control section 243 Communication circuit

Claims (13)

a memory that stores an observation matrix including a plurality of observation voltage vectors indicating a time series of cell voltages of each of the plurality of battery cells;
a control unit that determines a plurality of principal component vectors, a plurality of singular values, and a plurality of coefficient vectors from the observation matrix,
Each coefficient vector includes a plurality of coefficients corresponding one-to-one to the plurality of observed voltage vectors,
The control unit, for each coefficient vector,
comparing a plurality of coefficients included in the coefficient vector to determine an ineffective coefficient among the plurality of coefficients;
Based on the principal component vector corresponding to the coefficient vector among the plurality of principal component vectors, the singular value corresponding to the coefficient vector among the plurality of singular values, and the non-effective coefficient, the non-effective coefficient among the plurality of battery cells is A battery diagnostic device that detects abnormalities in battery cells corresponding to effectiveness coefficients. The control unit includes:
applying a matrix decomposition algorithm to the observation matrix to determine a first sub-matrix, a second sub-matrix and a third sub-matrix;
the first sub-matrix includes the plurality of principal component vectors as column vectors,
the second sub-matrix includes the plurality of singular values as main diagonal components,
The battery diagnostic device of claim 1, wherein the third sub-matrix includes the plurality of coefficient vectors as row vectors. The control unit includes:
The battery diagnostic device according to claim 1 or 2, wherein, among the plurality of coefficients, a coefficient whose absolute value of a difference from an average of the plurality of coefficients is larger than a first reference value is determined as the ineffective coefficient. The control unit includes:
The battery diagnostic device according to claim 3, wherein the first reference value is determined to be equal to the standard deviation of the plurality of coefficients multiplied by a first scaling factor. The control unit, for each coefficient vector,
Multiplying the principal component vector corresponding to the coefficient vector among the plurality of principal component vectors, the singular value corresponding to the coefficient vector among the plurality of singular values, and the ineffective coefficient, Extract the partial voltage vector of the observed voltage vector corresponding to the ineffective coefficient,
5. If the voltage characteristic value of the partial voltage vector is larger than a second reference value, a battery cell corresponding to the ineffective coefficient among the plurality of battery cells is detected as abnormal. Battery diagnostic device as described in. The control unit includes:
The battery diagnostic device according to claim 5, wherein the voltage characteristic value is determined to be the same as a difference between a maximum partial voltage and a minimum partial voltage among a plurality of partial voltages included in the partial voltage vector. The control unit includes:
The battery diagnostic device according to claim 5 or 6, wherein the second reference value is determined to be equal to the voltage resolution of the voltage measurement circuit multiplied by a second scaling factor. The control unit includes:
8. The battery diagnostic device according to claim 1, wherein a fault message is output when a ratio of a maximum singular value to a minimum singular value among the plurality of singular values is less than a set value. A battery pack comprising the battery diagnostic device according to any one of claims 1 to 8. A battery system comprising the battery pack according to claim 9. A battery diagnostic method, the method comprising:
determining a plurality of principal component vectors, a plurality of singular values, and a plurality of coefficient vectors from an observation matrix including a plurality of observed voltage vectors indicating a time series of cell voltages of each of the plurality of battery cells;
Each coefficient vector includes a plurality of coefficients corresponding one-to-one to the plurality of observed voltage vectors,
In the battery diagnosis method, for each coefficient vector,
comparing a plurality of coefficients included in the coefficient vector to determine an ineffective coefficient among the plurality of coefficients;
Based on the principal component vector corresponding to the coefficient vector among the plurality of principal component vectors, the singular value corresponding to the coefficient vector among the plurality of singular values, and the non-effective coefficient, the non-effective coefficient among the plurality of battery cells is detecting an abnormality in the battery cell corresponding to the effectiveness factor;
Further including battery diagnostic methods. The step of determining an ineffective coefficient among the plurality of coefficients includes:
12. The battery diagnosis method according to claim 11, further comprising determining, as the ineffective coefficient, a coefficient whose absolute value of difference from an average of the plurality of coefficients is larger than a first reference value, among the plurality of coefficients. Detecting an abnormality in a battery cell corresponding to the ineffectiveness coefficient among the plurality of battery cells,
Multiplying the principal component vector corresponding to the coefficient vector among the plurality of principal component vectors, the singular value corresponding to the coefficient vector among the plurality of singular values, and the ineffective coefficient, extracting a partial voltage vector of the observed voltage vector corresponding to the ineffective coefficient;
12. The method further comprises the step of detecting a battery cell corresponding to the ineffective coefficient among the plurality of battery cells as abnormal when the voltage characteristic value of the partial voltage vector is larger than a second reference value. Battery diagnostic method described in.

Images